| |
|
|
miqpBB
The solver MIQPbb
solves
large, sparse or dense
mixed-integer linear and
quadratic programming problems.
Main features
-
MIQPbb
implements a branch-and-bound tree search algorithm.
A new branching rule
has been proposed which uses lower bounds to decide on the
next fractional variable to branch upon.
The new branching rule has been shown to improve convergence.
The QP subproblems are solved using BQPD.
-
MINLPbb
may use three different branching rule options,
one being the above mentioned new rule.
-
The dense and the sparse version of
MIQPbb
are compiled in two different Mex binaries,
making the two versions optimally efficient.
-
MIQPbb
is integrated with the Tomlab graphical user interface
and the Tomlab driver routines.
-
MIQPbb
may be used as subproblem solver in the Tomlab
environment.
-
More information about
MIQPbb
is available at the
home pages of
Sven Leyffer,
University of Dundee.
|
|
